Demo Entry — This is a demonstration project used to preview the Open Research platform before launch. It does not represent a real, formally published research claim by Bacon Magic. Project contents are placeholders and will be replaced by reviewed, real projects.
Prototype

Camera-Based Dice Recognition

A prototype exploring camera-based dice recognition for mentalism applications. The system reads pips in real time and transmits the result discreetly to the performer.

Project ID
BM-OR-0001
Author
Bacon Magic R&D
Research Published
Oct 30, 2026
Version
v1.0
Disclosure
P2 - Method Disclosure

Research Overview

This project explores whether an off-the-shelf camera module can recognize dice values reliably enough for mentalism work, without custom electronics.

The goal is a system that reads a thrown die within one second and communicates the result to the performer through an existing wearable display.

Compared with known Dice-It style approaches that rely on mechanical or electronic dice, this design leaves every prop examinable: the dice are ordinary.

Inspiration & Development

Original inspiration: classic dice reading routines and modern camera-based prediction systems.

Our contribution: a pip-recognition pipeline that tolerates lighting changes and hand-thrown dice, using a low-cost ESP32-S3 camera module.

Related prior work: electronic dice readers sold by several magic dealers; we credit their pioneers for proving audience interest in this effect family.

Method & Technical Details

Show Method

The prototype uses an ESP32-S3 with an OV2640 camera. Frames are downsampled and thresholded on-device; pip positions are counted with a simple connected-component pass.

Current recognition accuracy is about 92% under typical indoor lighting. Failure cases are strong shadows and reflective table surfaces.

Status: working breadboard prototype. No PCB design yet.

Results & Known Limitations

Known limitations: (1) accuracy drops under stage lighting; (2) recognition distance currently limited to 40 cm; (3) battery life around 90 minutes.

Permission & Usage

M0 - Published for Reference

Published for study and reference. You may read, cite and discuss the published material. Reproduction of the prototype and any commercial development require separate written permission from Bacon Magic.

References & Credits

We distinguish original inspiration, our own contribution, and related prior work. Attribution we could not verify is marked as Unverified.

We acknowledge earlier electronic dice reading products in the magic market. Historical references will be added as they are verified.

Research History

  1. Oct 30, 2026
    First published on Open Research

    Initial v1.0 release of the project record.

Contributions & Disputes

Know earlier related work? Found an error? Have an improvement? Submissions are reviewed before anything is published, and your email is never shown publicly.

Submit a Contribution or Correction

How to Cite This Research

Bacon Magic. (2026). Camera-Based Dice Recognition. Bacon Magic Open Research. BM-OR-0001. https://baconmagic.shop/pages/bm-or-0001